Mr. Chairman, I appreciate very much the invitation to testify today about market power and structural change in the software industry. The question of the appropriate role of antitrust enforcers in high-technology industries has certainly been brought to a head with the government's recent enforcement action against Microsoft. And the resolution of this question is critically important to the future development of the software industry. These hearings therefore may serve to provide a timely and critically needed public forum to address this important question.
It is my own view that existing antitrust laws are entirely appropriate and sufficient to address anticompetitive behavior in software markets. It is critical, however, that these laws be enforced quickly and decisively to address anticompetitive behavior. I do not believe in excessive government regulation, and in particular I do not believe that the government should be in the position of picking winners or otherwise micro-managing the software industry or particular software businesses. Such intervention would crush the innovation that naturally results from open and unregulated competition. And I do not believe that such intervention is called for in today's software industry. However, if existing antitrust laws are not vigorously enforced today, I fear that the resulting monopolization of various software markets will lead inevitably to calls for such government regulation.
Sun Microsystems, Inc. ("Sun") has been challenging the "mainstream" computer industry and forging new technologies since the company was founded sixteen years ago. Sun was started by four individuals in their mid-twenties who envisioned a computer that could sit on the desktop and provide a level of computing power similar to much larger machines at a fraction of the cost. Importantly, Sun's founders, myself included, believed that Sun computers must be based on open standards. This then-unique concept would allow customers to mix and match Sun systems with products made by other vendors. Customers no longer would be locked into buying proprietary technologies made by one company.
Sun's corporate vision to make the maximum computing power available to a broad user audience through open technologies has not changed since the company's inception. What has changed is the computing environment and the market. Sixteen years ago, computing power was often centralized within companies on mainframes and minicomputers. The actual computers available to individual users were, in fact, "dumb" terminals, or stand alone, low performance PCs. The high performance computers available at the time were prohibitively expensive. These factors helped make Sun's networked workstations so successful.
Today, Sun helps large and medium sized companies overcome the limitations of their legacy mainframe and minicomputer systems. Sun is also helping to lead the way into the next generation of computing Internet and network-centric computers. Sun's focus on network computing is evident by our slogan "The network is the computer." Sun's Java technology, described in more detail below, is one indication of our commitment to open network computing.
We have recently entered the third revolution in computing. The first, dominated by IBM, was the development of the mainframe computer and centralized processing. The second, dominated by Microsoft, IBM's inadvertently anointed successor, was the switch to PC desktop computing. The third and current revolution is the rise of Internet and network computing based on competition and innovation enhancing open standards. The essential question posed to antitrust enforcers today is whether dominant firms in PC desktop markets should be allowed to leverage their monopoly power through anticompetitive and exclusionary practices to extend their dominance from the second to the third revolution. I believe that such an outcome would necessarily undermine the open and free standards which have spawned this third revolution and fueled its rapid growth, and therefore must be prevented.
The distinction between the open standards of Internet and network computing versus the closed standards of mainframe and PC computing is critical, particularly as it impacts innovation. The mainframe computer industry of the 1950s, 60s and 70s was dominated by IBM, an integrated company which sold mainframes, and later, minicomputers based on proprietary technology. Because its equipment would not work with products made by other companies, customers essentially were forced to rely on a single firm for all of their service and technology solutions. As a result, IBM's technical standards ruled the industry.
Due to a strategic error on the part of IBM, its dominance was lost in the PC era. Specifically, rather than produce software technologies on which this new PC industry would be based, IBM contracted this work out and focused instead on maintaining control over the boxes to house the technology. By failing to see the strategic importance of controlling the operating system software standards, IBM unwittingly ceded control of the PC operating systems market to Microsoft which owned the underlying intellectual property. Thus, when IBM decided to anoint Microsoft as the sole supplier of the IBM PC operating system, IBM inadvertently transferred its market power to Microsoft. In the fifty-year history of this industry there have primarily been two dominant firms and domination by one led directly to domination by the other. This domination was and is maintained in large part through control of proprietary standards. There is one huge difference: when IBM dominated the industry, information technology played a far smaller role in the economy and in the lives of our people than it does today. As Microsoft seeks to leverage its PC operating system and application monopolies up into the enterprise, down into consumer products, and out into the Internet, it has the power to affect the economy, and the way we live and work, that IBM, even at its zenith, never had.
Today, 85% of the world's existing PCs contain operating systems from Microsoft and over 95% all PCs currently being shipped contain a Microsoft operating system. Currently, there is virtually no competition in the PC operating system market, and PCs with Microsoft operating systems are now the gateway device by which the overwhelming majority of the consumers access the Internet. However, the rapid rise of the Internet now threatens, at least potentially, to undermine existing monopoly power in the PC market by enabling rapid innovation and development of new software under open standards.
Understanding the critical importance of standards is the key to understanding the future of competition in the software industry, and is related to the existence of "network effects" in this industry. In a market characterized by network effects, the value of a product to existing users increases with each additional user. Such increasing returns in consumption generate a "positive feedback" effect: the more users a product garners, the more attractive it becomes to other customers. Large user bases are more attractive to new customers, and thus get larger. Sufficient positive feedback can result in "tipping" the market toward the product or standard that has received the widest adoption, whether or not it embodies the most superior technology. Once tipping occurs, "lock-in" of customers may result, such that switching subsequently to a superior product or technology may prove too costly for users. These costs result from sacrificing compatibility with other users, discarding a significant investment in complementary products, or incurring new learning costs.
Another factor distinguishing software markets from traditional manufacturing industries is the tremendous economies of scale inherent in the production of software. Virtually all the costs of software production are incurred in research and design of the software, which are fixed costs. The marginal cost of software production is almost zero, entailing essentially the cost of making copies of the software, and maybe also, the cost of printing a manual.
Network effects combined with economies of scale in production can cause software markets to tend toward single firm dominance. The "first-mover" in many software markets may have a significant advantage in the race to tip the market in favor of its technology or standard, garner an installed base of users, and establish an insulated, dominant position for itself. A first-mover may sometimes even give away its product in order to establish a sufficient installed base to tip the market in its favor and then charge for further upgrades to its products.
The open standards that have developed with the emergence of the Internet and other network-centric technologies have the potential to threaten monopoly power in PC software markets by reducing the network effects entry barriers that new operating systems and application software must overcome. Sun's Java technology and Netscape Communication Corp.'s ("Netscape") Communicator and Navigator software ("Netscape technology") are clear examples of these emerging open standards. Java technology provides a means for creating applications that can not only be run on standalone PCs, but can also be transferred over the Internet (or any other network) and then run on any platform supporting the Java standard without the need to rewrite the application for each platform. Java technology's platform-independence property minimizes the importance of any particular operating system thus freeing application developers to write applications without the constraint of having to write for multiple operating systems "write once, run anywhere." It also allows consumers the freedom to select applications without the added constraint of selecting a platform. Java technology therefore eliminates or significantly reduces the network effects entry barrier that new operating systems must overcome. In the pre-Java world, a new operating system standing alone would face insurmountable difficulties in successfully entering the marketplace, even if it was technically superior to existing operating systems. Such a system would be hampered by the lack of a full set of applications already written for it, and by the fact that users are already locked into the Windows technical standards. If platform-independent Java-based applications become widely available, consumer choice in operating systems and applications will be increased.
Netscape technology also poses a threat to current dominance in PC markets because it can offer consumers a platform-independent Web-based user interface ("WUI"). WUIs are essentially "virtual desktops" that can overlay existing PC desktops much the same way the Microsoft Windows desktop replaced the older DOS interface.
It appears that Microsoft's principal weapons in maintaining its PC operating system monopoly in the face of these threats has not been the promotion of better more innovative products through competition, but rather the exclusion of new technologies, including Java and Netscape technologies, through anticompetitive practices that unfairly favor Microsoft's proprietary version of Java technology and Microsoft Internet Explorer ("IE"). The recent Antitrust Division action against Microsoft involves an allegation of the erection of one of these artificial entry barriers by Microsoft, not the government, in order to maintain and extend its dominance. Specifically, by forcing OEMs to license IE as a condition of licensing Windows 95, the Antitrust Division alleges that Microsoft is attempting to use its operating system monopoly to achieve an unfair advantage over competing Internet browsers. Other alleged examples abound of Microsoft's erection of artificial barriers, i.e., conduct intended simply to thwart competition as opposed to being efficiency based:
Under its Windows licensing agreement, Microsoft allegedly requires computer OEMs to configure their PCs so that a Microsoft desktop appears upon boot-up. This requirement appears designed to prevent Netscape technology from directly competing with Microsoft by prohibiting OEMs from replacing the Windows desktop with the Netcaster desktop upon boot-up.
Microsoft allegedly signed agreements with the major online services and Internet service providers ("ISPs") whereby these providers offer IE as their exclusive browser. In exchange, Microsoft offers direct access to these providers in Windows through an "online services" desktop folder.
Microsoft allegedly signed agreements with more than 250 content providers to deliver content to the IE 4.0 "Active Desktop" channels using push technology. As part of these agreements, Microsoft is "insisting its partners exclusively promote Internet Explorer, limiting their ability to work with Netscape." According to one content provider executive, "Microsoft is saying If you're with us, you can't be in with somebody else.'"
Left unchecked, the long list of artificial barriers allegedly erected by Microsoft will lead to the replacement of open standards, such as those embodied in Netscape and Java technologies, with Microsoft's proprietary platform-dependent Windows standards. The essential question posed to antitrust enforcers is whether dominant firms in standalone PC markets, such as Microsoft, should be allowed to extend their dominance to emerging Internet and network-centric markets by unlawfully undermining the open standards which have fueled the rapid growth of the Internet through exclusionary acts designed to impose their own proprietary standards in these emerging markets.
Whereas in traditional industries price is king, and consequently the central focus of antitrust analysis, in software markets innovation is at least an equivalent driver of competition. Perhaps the most vivid current illustration of this principle is the proliferation of new companies with new and innovative technologies that have recently sprung up due to the rising importance of the Internet. Indeed, in these emerging markets it is common practice for firms to initially give their products away as a means of tipping the market in their favor. As far as antitrust law is concerned, the competitive issues presented by market power in software markets are not new. Courts have long recognized in applying antitrust laws that dynamic efficiency is critical to the long-term growth of consumer welfare, and have repeatedly cautioned that antitrust enforcement must not "chill" or "stifle" innovation. Thus, the Supreme Court has held that once exclusionary behavior is identified, antitrust liability will turn on whether such conduct is supported by "valid business reasons." The antitrust enforcement agencies recognize the significance of innovation as a driving market force, as indicated by the attention to innovation markets in the 1995 intellectual property guidelines. The overriding objective of these guidelines is to ensure a free market for innovation, with the benefits that flow therefrom, thereby maximizing consumer welfare.
Nowhere is this well accepted insight more apparent to antitrust enforcers than in the rapidly evolving computer industry. As stated by then-Deputy Assistant Attorney General Carl Shapiro:
It is not infrequent for one firm to wrest industry leadership away from another as technology advances from one generation to the next. This is Schumpeterian creative destruction' at work to deliver ever-better products to consumers. The single most important goal of antitrust in network industries is to insure that competition from new products and new technologies is not stifled.
. . . .
. . . Having worked with dynamic, game-theoretic models of business strategy for my entire professional career, I am well aware of the pitfalls of employing static analysis in dynamic industries, and the information industries are nothing if not dynamic.
In applying existing antitrust laws to software markets today, antitrust enforcers must recognize innovation as the driving force in software markets. In particular, enforcers must strive to avoid two market outcomes in software industries that are antithetical to innovation. The first is market monopoly and the second is market regulation. Both outcomes must be avoided if emerging Internet markets are to thrive. As a successful business person in emerging software markets, I am well aware of the importance of competition in spurring innovation. Firms that fail to innovate simply will not survive. The protection that monopoly provides simply eliminates this incentive to innovate. Indeed, in the highly monopolized PC operating systems market innovation has been minimal, a trend strikingly absent among other software markets.
In industries such as software where network effects are so prevalent, the ability of monopolists to stifle innovation is even greater. Where network effects are present, dominance can occur as the result of natural market forces, and the role of antitrust enforcers should be limited. Certainly, antitrust enforcers or other government agents should not attempt to pick winners. However, the existence of natural entry barriers to would-be entrants/innovators should cause antitrust enforcers to be all the more vigilant in attacking the erection of artificial entry barriers by a dominant firm. However, it will often be the case that unique opportunities will present themselves to the firm that has achieved monopoly power to engage in exclusionary behavior to maintain that existing monopoly power, as well as to attempt to garner monopoly positions in complementary or related markets. It may also be the case that a dominant firm seeking to leverage its position will engage in every imaginable twist to characterize the artificial barriers being erected as natural barriers whose presence is inevitable. Those assessing the need or propriety of government enforcement should be alert to the presence of such sleights of hand. These opportunities usually involve foreclosing innovation directly by maintaining control of proprietary standards, excluding new innovators by preventing access to closed systems, or by acquiring the upstart competitor outright thereby eliminating any possibility that the competitor may pose a future threat. Such conduct is antithetical to consumer welfare because it is not motivated by efficiency or innovation concerns, but simply is intended to protect the monopoly position. This problem is well recognized by economist and antitrust enforcers. Nobel laureate Kenneth Arrow outlined the issue as follows:
The history of market shares in PC application software has been marked by great volatility. Although first-mover advantages and increasing returns are important, there are many examples to show that such advantages are far from permanent. As examples, consider the fates of Wordstar, Apple Computer, and IBM itself. All were once dominant in critical PC-related product markets; yet each has experienced rapid loss of market shares.
Noting this transience is not justification for complacency. On the contrary, it requires effort to maintain the openness of markets, so that new technologies can have an opportunity to enter and show their value relative to older ones.
Then-Deputy Assistant Attorney General Carl Shapiro has also commented on the interplay between technological change, network effects and entry barriers in high-technology markets:
Because innovation is such a strong force in many high-technology markets, companies are tempted to defend their conduct by arguing that entry is easy or inevitable, and thus durable market power or monopoly power is unobtainable. Sometimes this argument may be quite valid, but beware of overusing it: there is no antitrust immunity for high-tech industries. The fact is, rapid technological progress does not equate to low entry barriers, especially if users find it very costly to switch to new brands that are incompatible in some way with the established technology.
. . . .In some cases, the leader in one generation of technology is able to perpetuate its dominance into the succeeding generation by offering the best technology to users; this represents healthy competition. But antitrust concerns quickly arise when a firm controlling the standard in one product area uses its dominance to set and control the standard for the next generation of that product, or for a second, complementary product.
With regard to the second market structure that is antithetical to innovation government regulated markets, one need look no further than to the failed history of government regulation in airlines, telephones and electric utilities to see that excess regulation does not foster innovation. First, with regulation comes the death of competition and hence the elimination of the need to innovate for survival. Second, time and time again history has shown that the regulated ultimately capture the regulators to preserve their government-anointed monopoly, thereby eliminating the possibility of innovation from new entrants.
In order to avoid these two market outcomes, antitrust enforcers must vigorously and rapidly apply existing antitrust laws to any and all anticompetitive behavior that has no justification but to foreclose innovation and competition. Not to do so poses two risks. First, emerging Internet and network markets based on open platform-independent technologies, such as Netscape and Java technologies, will be supplanted by existing closed standards now prevalent in PC software markets. Second, once PC software firms extend their dominance to emerging Internet markets, there will be no unscrambling the eggs. At that point, proponents of government regulation may have their way, and a bad situation will inevitably become worse as government regulators attempt to micro-manage the software industry. Based on the above concerns, I would like to end by summarizing my view regarding what antitrust enforcers should and should not do.
What antitrust enforcers should not do:
Antitrust enforcers should not attempt to pick winners. Competitive forces that play out on a level field are always the best determinants of winners and losers.
Antitrust enforcers should not attempt to micro-manage the software industry or any particular software business. Such regulators will inevitably by captured and manipulated by the regulated.
Antitrust enforcers should not punish dominance gained through efficiencies and open competition. Monopoly itself is not illegal, it is the attainment, maintenance and extension of monopoly through anticompetitive and exclusionary practices that is illegal. As noted above, in industries such as software, where network effects are prevalent, the existence of natural entry barriers to would-be entrants/innovators should cause antitrust enforcers to be all the more vigilant in attacking the erection of artificial entry barriers by a dominant firm.
What antitrust enforcers should do:
Antitrust enforcers should act rapidly and vigorously to prevent the maintenance and extension of monopoly power through exclusionary practices such as the erection of artificial barriers or anticompetitive acquisitions designed to block efforts by would-be entrants to offer innovative technologies. All to often, past efforts at enforcement have been too little, too late, as the targeted anticompetitive behavior has already achieved its goal. In short, when a decision to bring an enforcement action is made, act vigorously and quickly.
Antitrust enforcers should be particularly vigilant of the following exclusionary practices: (1) efforts to replace emerging open standards with existing dominant closed standards; (2) efforts to leverage dominance into related markets through exclusionary practices; (3) efforts to prevent access to OEMs and ultimately to the PC desktops of consumers; and (4) efforts to forestall innovation through anticompetitive acquisitions of innovative startup firms.